Mapping critical events with UAV-based photogrammetry

Precise mapping plays an important role when documenting critical events such as car accidents or crime scenes. In contrast to terrestrial mapping, UAVs allow taking images from above the event, thus yielding bird’s eye views of the scenario at photogrammetric accuracies. The research and development is based on ground control markers, i.e. metal plates with custom designed optical markers, which are positioned at critical locations and measured with high accuracy using GNSS equipment by the police. These 3D coordinates serve as input for a UAV, such as a DJI Inspire 2, that can autonomously position itself centralized above the scenario and continuously take images. These images are processed by state-of-the-art computer vision techniques to detect the markers in the images (currently so-called ArUco markers are employed). The detected number and position of the markers are used to navigate the UAV to a best fitting location and orientation in space where one or more very high resolution images are taken covering the whole event. Since 3D world coordinates of the markers are known and their 2D image coordinates respectively, the image(s) can be georeferenced either on a plane or on a given digital surface model. This metric information is stored in a database where certain features, like skid marks, can be analyzed even a long time after the event, which is one very important aspect for law suits and in particular for insurance companies. In addition, several stereo images from the event will be recorded in order to allow a precise 3D reconstruction. The system is embedded into the integrated police system (iPolSYS) that is used by the Austrian police.